AASHTO Petition For Reconsideration Of The 4.9 GHz 7th Report And Order
AASHTO Comments on FCC’s Ninth Further Notice of Proposed Rulemaking in the 4.9 GHz docket
Re-considering 5.9 GHz spectrum sharing
by Julian Gehman
Gehman Law PLLC, Washington, DC, USA
Abstract
This paper reviews key assumptions underlying the initiative of the Federal Communications Commission (FCC) to share, with unlicensed national information infrastructure (U-NII) devices, the 5.9 GHz spectrum band that is allocated to dedicated short range communication (DSRC). These assumptions are: (1) that there is a shortage of unlicensed spectrum available to U-NII devices, and the 5.9 GHz band is vital to deployment of high-speed Wi-Fi and 5G, (2) that similar incumbents across the 5 GHz band would facilitate harmonized rules for unlicensed use at 5.9 GHz, and this in turn would allow for economical U-NII devices and more efficient use of aggregated spectrum, and (3) that 5.9 GHz spectrum-sharing was mandated by President Obama’s Presidential Memorandum and the Spectrum Act. In summary, these assumptions are misplaced: (1) There is no shortage of unlicensed spectrum for U-NII devices, but much inefficient use. Continued use of a spectrum-inefficient Wi-Fi configuration is mostly responsible for the projected shortage of unlicensed spectrum at 5 GHz. Nevertheless, the FCC is conducting a large spectrum allocation that should more than satisfy reasonably projected demand, even with the inefficient configuration; (2) DSRC is very different from the other incumbent services in the 5 GHz band, requiring different protective measures; and (3) neither the Presidential Memorandum nor the Spectrum Act require 5.9 GHz spectrum sharing. For these reasons, the rationale for sharing the 5.9 GHz band seems questionable.
KEYWORDS
5.9 GHz, DSRC, U-NII
Introduction and summary
Proposed sharing of the 5.9 GHz spectrum band is one of the more serious regulatory impediments to deployment of dedicated short-range communication (DSRC). A National Academies of Science committee said that the FCC’s 5.9 GHz spectrum-sharing initiative poses “the most serious risk and uncertainty” to DSRC.[1] Similarly, Toyota Motor Company North America told the FCC that the company was deploying DSRC in Japan and “will be in a position to promptly deploy similar systems in the United States once the current uncertainty regarding the operation of DSRC systems at 5 GHz is resolved.”[2]
In 1999, the FCC allocated 75 MHz of spectrum at 5.85 GHz to 5.925 GHz for use by DSRC (the “5.9 GHz band”).[3] In February 2013, the FCC proposed that the 5.9 GHz band be shared with Unlicensed National Information Infrastructure (U-NII) devices.[4] However, the Institute of Electrical & Electronic Engineers (IEEE) – author of the 802.11a, 802.11n and 802.11ac standards used by U-NII devices, and also author of the 802.11p standard used by DSRC – could not reach agreement on how to get these two sets of standards to co-exist in the 5.9 GHz band. Two competing protocols, re-channelization and detect-and-avoid, were presented, with the IEEE Tiger Team failing to reach consensus.[5] The FCC subsequently announced a plan to test prototypes of U-NII and DSRC devices, to determine whether they could co-exist in the 5.9 GHz band without causing harmful interference to DSRC service.[6] The FCC’s plan included Phase I, Phase II and Phase III testing and was to be completed and submitted by January 15, 2017.[7] According to informal reports, testing has been delayed by a dearth of prototypes, among other reasons. As of the writing of this paper (March 2018), the FCC has completed its Phase I testing, but has not reported any results. The FCC’s planned Phase II and Phase III tests presumably are upcoming, and the agency’s study apparently could go on for several more years. Lack of clarity about spectrum sharing, which has impeded DSRC for five years, is set to continue into the foreseeable future.
With the passage of time, the assumptions behind this initiative should be re-examined. This paper provides a critical review of the claim that the 5.9 GHz band must be shared with U-NII devices. In particular, this paper examines three key assumptions that have driven the 5.9 GHz spectrum-sharing initiative: (1) that there is a shortage of unlicensed 5 GHz spectrum, and the 5.9 GHz band is vital to deployment of high-speed Wi-Fi and 5G, (2) that similar incumbents across the 5 GHz band would facilitate harmonized rules for unlicensed use at 5.9 GHz, and this in turn would allow for economical U-NII devices and aggregated spectrum for high-speed Wi-Fi, and (3) that 5.9 GHz spectrum-sharing was mandated by the Spectrum Act and President Obama’s Memorandum of June 28, 2010.
In summary, these assumptions are misplaced, and the rationale for sharing the 5.9 GHz band seems questionable. (1) The 5.9 GHz band is not needed for 5G or gigabit Wi-Fi. Continued use of a spectrum-inefficient Wi-Fi configuration is largely responsible for the projected shortage of unlicensed spectrum at 5 GHz. The focus should be on improving spectrum efficiency rather than allocating more unlicensed spectrum to be used inefficiently. Nevertheless, existing unlicensed spectrum at the 5 GHz band is sufficient to meet current demand, and additional allocation in the FCC’s pending Mid-Band Proceeding should provide a substantial amount of unlicensed mid-band spectrum that should more than satisfy the projected needs of the inefficient Wi-Fi configuration, all without using the 5.9 GHz band. (2) DSRC is different from other incumbents in the 5 GHz band. DSRC is much less tolerant of radio interference. Under a detect-and-avoid regime, any U-NII device operating at 5.9 GHz would have to employ new equipment, including a new chipset, and would be subject to stringent rules which, from a U-NII operator’s perspective, could be worse than dynamic frequency selection (DFS). Finally, (3) the Spectrum Act mandates a study, not re-allocation of the 5.9 GHz band, and President Obama’s Memorandum, calling for 500 megahertz of spectrum to be made available for broadband, has been more than satisfied by other allocations. These points are reviewed below.
Background on U-NII devices
U-NII devices provide wireless Internet access and other wireless services. U-NII devices can be mounted in a public place above roof-top level, providing point-to-multipoint service. For example, the popular Canopy U-NII device, manufactured by Motorola, Inc., has a range of approximately two miles unassisted, or ten miles with reflectors.[8] However, the most common U-NII device is the Wi-Fi “hot spot” or access point found in residences and businesses. According to the FCC, “U-NII devices are unlicensed intentional radiators, which use wideband digital modulation techniques to provide a wide array of high-data-rate mobile and fixed communications used by individuals, businesses and institutions, particularly for wireless local area networking – including Wi-Fi – and broadband access.”[9]
The bigger the bandwidth of U-NII transmission, the faster the speed. U-NII devices operating at 5 GHz utilize three Wi-Fi standards with increasingly greater bandwidths and speeds: IEEE 802.11a (20-megahertz channel bandwidth and maximum date rate of 54 Mbit/s), 802.11n (20- or 40-megahertz bandwidth, with data rates up to 600 Mbit/s on the 40 megahertz channel), and 802.11ac (20-, 40-, 80- and 160-megahertz bandwidths, with data rates up to 1Gbit/s). Finally, a new standard, IEEE 802.11ax is expected in 2019. Wi-Fi providers seek to deliver “gigabit Wi-Fi” for higher capacity downloads and streaming. Gigabit Wi-Fi is supposed to provide better user experience and facilitate battery conservation. According to cable television companies, 80- and 160-megahertz bandwidths of 5 GHz band spectrum are needed in order to provide gigabit Wi-Fi. However, as discussed below, high-band (60 GHz) spectrum is better suited for gigabit Wi-Fi.
In 1997, the FCC made available a total of 300 megahertz of spectrum, for use by U-NII devices, at 5.15-5.25 GHz (referred to as the U-NII-1 band), 5.25-5.35 GHz (U-NII-2A), and 5.725-5.825 GHz (U-NII-3). In 2003, the Commission made available an additional 255 megahertz of spectrum, for use by U-NII devices, at 5.47-5.725 GHz (U-NII-2C). In 2014, the FCC extended the upper edge of the U-NII-3 band by 25 megahertz (U-NII-3 band now is 5.725-5.850 MHz), bringing the total amount of 5 GHz spectrum, currently allocated to U-NII use, to 580 megahertz.[10] The Commission stated that “[t]hese actions align the frequency bands used by U-NII devices in the United States with the frequency bands used by U-NII devices in other parts of the world, thus decreasing development and manufacturing costs by allowing for the same products to be used in most parts of the world.”[11]
U-NII devices operate on a secondary basis to incumbent services. In the 5 GHz band, the incumbents mostly are radar and satellite uplink. As a secondary user, a U-NII device, must accept whatever interference is received and must not cause harmful interference. Radar and satellite uplink stations generally are stationary and their locations are publicly known. Thus, U-NII devices can at least partially avoid interfering with these stations using informal geolocation fencing. In addition, U-NII devices operating in the U-NII-2A and U-NII-2C bands must employ Transmit Power Control (TPC) and Dynamic Frequency Selection (DFS).[12] With TPC, higher power U-NII devices must reduce power. With DFS, the U-NII device must listen to detect a radar signal, both before and during U-NII transmission. Prior to initiating a transmission, the U-NII device must listen for a radar signal for 60 seconds. After a radar’s presence is detected, all U-NII transmissions must cease on that channel within 10 seconds, and the channel is then subject to a non-occupancy period of 30 minutes.[13] Indoor Wi-Fi, using lower power, is subject to a relaxed DFS sensing requirement when operating in the DFS-limited bands[14] and is less susceptible to disruption from DFS.
The Commission subsequently harmonized power and use conditions across the lower 200 megahertz of U-NII spectrum (U-NII-1 and U-NII-2A) in order to “permit the introduction of a wide range of new broadband products capable of operating at higher data rates.”[15] In particular, this harmonization effectively created 200 megahertz of contiguous spectrum that could be used for 80- and 160-megahertz bandwidths under the IEEE 802.11ac standard. There are two 160-megahertz bandwidths currently available in the 5 GHz band: (1) the combined U-NII-1 and U-NII-2A bands, discussed above, and (2) the U-NII-2C band, which has 255 megahertz of contiguous spectrum. Both of these 160-megahertz channels are subject to TPC and DFS which hamper outdoor, higher power Wi-Fi but have less effect on indoor, lower power Wi-Fi.
Finally, the Consolidated Appropriations Act, 2018, signed into law by President Trump on March 23, 2018, includes the MOBILE NOW Act, which impacts the FCC’s allocation of unlicensed spectrum.[16] Section 617 of the MOBILE NOW Act makes it the policy of the United States provide “on an unlicensed basis radio frequency bands to address consumer demand for unlicensed wireless broadband operations.” Section 618 of the same statute requires the FCC to develop a national plan to provide consumers with unlicensed frequencies in various bands and steps to identify additional unlicensed spectrum or to use existing unlicensed spectrum more intensively.
First assumption: shortage of unlicensed spectrum requires use of the 5.9 GHz band
The FCC uses auctions to assign licensed spectrum to the parties who most highly value the spectrum.[17] Congress authorized the FCC to conduct auctions after spectrum lotteries were shown to give a windfall to parties that had no intention of putting the spectrum to constructive use.[18] Spectrum auctions serve as a market mechanism for spectrum allocation and a market discipline to ensure that spectrum is used efficiently. Similarly, the FCC imposes construction requirements to ensure that, after a certain period of time, usually ten years, licensed spectrum is put to use.[19] The FCC reclaims spectrum that has not been the subject of timely construction and use. This too serves as a market mechanism to inform the Commission of spectrum usage and to ensure that the spectrum is put to use.
Of course, the point of unlicensed spectrum is that it is not auctioned, not licensed and not subject to construction requirements. As a result, the FCC’s decision to allocate unlicensed spectrum is untethered from market mechanisms. The FCC has not published any projection of the amount of spectrum that could be needed for unlicensed use, nor does the agency measure the degree of usage of unlicensed spectrum. The FCC also has no measure of whether unlicensed spectrum is being used efficiently. The FCC relies on reports and forecasts from parties that have an economic interest in greater allocation of unlicensed spectrum. Two such forecasts have gained prominence: Wi-Fi Alliance’s “Wi-Fi Spectrum Needs Study” (the “Wi-Fi Alliance Study”),[20] and Qualcomm’s “A Quantification of 5 GHz Unlicensed Band Spectrum Needs” (the “Qualcomm Study”).[21] These two studies arrive at similar conclusions as to the amount of mid-band spectrum that is needed. The Qualcomm Study also demonstrates that Wi-Fi service is being delivered in a spectrum-inefficient manner, and if spectrum-efficient configurations were utilized, there would be no need to allocate additional mid band spectrum. These studies are briefly reviewed below.
The Wi-Fi Alliance Study
The Wi-Fi Alliance Study is a bottom-up projection of wireless traffic, derived from three sources: Cisco VNI Forecast Tool, an Offcom UK study and a commercial study. According to the Wi-Fi Alliance Study, “[t]he predicted demand from the three independent sources is reasonably aligned and differences will likely not significantly affect the amount of spectrum required.”[22] The Wi-Fi Alliance Study presents a Busy Hour scenario and an Upper Bound scenario, and “the Busy Hour scenario is the most likely to occur while the Upper Bound scenario is less likely but still plausible.”[23] The term “Busy Hour” means that radio traffic projections were studied for the busiest four hours of the day (7:00 pm to 11:00 pm), because a network needs to be provisioned for the period of greatest usage.[24] The Wi-Fi Alliance Study comes up with its Upper Bound scenario by taking the average traffic in the Busy Hour projection for 2025 and multiplying by four.[25] Frequency re-use and congestion are important factors in the Wi-Fi Alliance Study.[26] The Wi-Fi Alliance Study emphasizes indoor consumer use, noting that outdoors use is one tenth the volume of indoors use, and office use is one quarter the volume of consumer use.[27] For residential use, the study assumes dense apartment buildings – each apartment has 1076 square feet, four occupants per apartment, and three devices per occupant, with multiple apartments per floor and multiple floors in the apartment building.[28] Based on surveys conducted in London and San Francisco, the Wi-Fi Alliance Study finds that channels subject to DFS are less heavily used. Consequently, the Wi-Fi Alliance Study includes separate projections for 30% of DFS spectrum in use, and 100% of DFS spectrum in use.[29]
The Wi-Fi Alliance study projects the total amount of spectrum needed, takes into account the existing unlicensed spectrum, and performs a gap analysis by subtracting existing spectrum allocated from total spectrum needed.[30] The result of the gap analysis is the amount of new, unlicensed spectrum that the Wi-Fi Alliance Study says is needed. The Wi-Fi Alliance Study concludes that, by the year 2025, under its Busy Hour scenario, a total of 1120 MHz of unlicensed spectrum around 5 GHz will be needed.[31] The United States has 580 megahertz of 5GHz spectrum allocated to unlicensed use. Therefore, according to the study, the United States will need 540 megahertz of new spectrum at 5 GHz if all DFS spectrum is in use, and 788 megahertz if 30 per cent of DFS spectrum is in use.[32]
Under the study’s Upper Bound scenario, the United States would need 1340 megahertz and 1588 megahertz, respectively, in 2025.[33] According to the study, the Upper Bound scenario seeks to account for situations where more spectrum is needed than was projected. In order to accomplish this, the study multiplies base traffic by four. No explanation is given for quadrupling rather than some other factor. With any projection, there is the possibility of contingencies. Decisions should be based on the most likely outcome, not on an arbitrary multiple to account for a remote contingency. As noted, the Wi-Fi Alliance Study says that the Busy Hour scenario is the most likely to occur. The study’s reasonable projection is that of the Busy Hour scenario – 540 megahertz of new, unlicensed spectrum if all DFS spectrum is in use and 788 megahertz if 30 per cent of DFS spectrum is in use.
The Qualcomm Study
The Qualcomm Study recommends that “[t]o enable future WLAN-type applications and usage scenarios, regulators need to plan for around 1280 MHz of unlicensed spectrum centered around the 5 GHz band” and that “regulators should strive towards making multiple (i.e., 3 or more) 160-megahertz wide channels available for unlicensed use.”[34] Qualcomm says its study is a “top-down, engineering driven, analysis of required spectrum to achieve ‘wired equivalent’ performance for unlicensed spectrum technologies in dense networking environments.”[35] The 1280 megahertz figure represents the total amount of spectrum that Qualcomm says is needed to get the job done: “The required spectrum is determined by finding the minimum amount of spectrum needed to achieve the target throughput rate” of 1 Gbit/s.[36]
The Qualcomm Study’s projection of total need of 1280 megahertz is consistent with the Wi-Fi Alliance Study’s projection of 1120 megahertz. However, unlike the gap analysis performed by the Wi-Fi Alliance Study, the Qualcomm Study does not account for spectrum already allocated. Nevertheless, the 580 megahertz already allocated to U-NII use in the 5 GHz band would contribute toward satisfying the overall requirement identified in the Qualcomm Study. Therefore, the Qualcomm Study is saying that 700 megahertz (1280 total need minus 580 already allocated equals 700 of remaining need) of new, unlicensed spectrum in the 5 GHz band is needed.
Comparison of forecasts with the FCC’s spectrum allocation
The projections of the two studies are consistent with each other. After performing a gap analysis, the Qualcomm study projects that 700 MHz of new, unlicensed spectrum around the 5 GHz band would be needed while, under the Busy Hour scenario of the Wi-Fi Alliance Study, 540 to 788 megahertz would be needed in 2025. Both studies emphasize the need for additional 160-megahertz channels.
In August 2017, the FCC initiated a proceeding to allocate a total of 1700 megahertz of spectrum for mobile broadband use (the “Mid-Band Spectrum Proceeding”).[37] The preliminary plan is to make 500 megahertz into auctioned, licensed spectrum, and the remaining 1200 megahertz at 5.925-7.125 GHz (the “6 GHz band”) into unlicensed spectrum for use on a secondary basis.[38] Were to Commission to allocate 1200 megahertz to U-NII use, or even something close to that amount, it would exceed by a large margin the need that is reasonably projected by the Qualcomm Study and the Wi-Fi Alliance study. Thus, there should be more than enough mid-band unlicensed spectrum without tapping into the 5.9 GHz band.
Incumbent licensees operate in this 1200 megahertz of spectrum, who should be protected. Therefore, some or all of the 1200 megahertz at 5.925-7.125 GHz would be subject to one or more sharing mechanisms. Recognizing the disadvantages of DFS, commenters to the Mid-Band Spectrum Proceeding have proposed alternative spectrum sharing mechanisms to protect incumbents without imposing DFS. The IEEE 802 Committee stated its confidence that these protections would be “as effective as DFS has been.”[39] Preliminarily, it appears that this 1200 megahertz contains enough contiguous spectrum to form three or more 160-megahertz bandwidths that are not subject to DFS.
Continued use of a spectrum-inefficient Wi-Fi configuration drives projections of mid band spectrum shortage
The Qualcomm Study models a variety of in-home Wi-Fi configurations, which result in wide variance of 5 GHz spectrum needed for each configuration. The Qualcomm Study models either two or four antennas at the device/client side (STA) for each of the following residential configurations, with the following results:
| Configuration Title | Residential Configuration For Delivery of Gigabit Wi-Fi | Amount of 5 GHz Spectrum Required (in Megahertz) |
| Config. A | 1 access point (AP) per apartment, STA – 4 Antennas | 1280 |
| Config. B | 4 APs per apartment with Ethernet backhaul | 320 |
| Config. C | 4 APs per apartment, WiGig 60 GHz last hop, 5 GHz WLAN backhaul | 480 |
| Config. D | 4 APs per apartment, 5 GHz WLAN last hop, 5 GHz WLAN backhaul, STA – 4 Antennas | 800 |
Table 1 Spectrum efficiency of residential access point configurations. Source: Qualcomm Study
Configuration A requires 1280 megahertz of 5 GHz spectrum to provide gigabit Wi-Fi to a four-room apartment, while Configuration B requires 320 megahertz. In other words, Configuration A requires four times as much 5 GHz spectrum to do the same thing as Configuration B. This large disparity demonstrates the profound effect that choice of residential access-point configuration has on the amount of 5 GHz band spectrum that is needed.
Configuration B (320 megahertz) or Configuration C (480 megahertz) apparently could be accomplished with no further allocation of 5 GHz unlicensed spectrum beyond the 580 megahertz already allocated. A configuration of four APs with Ethernet fronthaul and backhaul and 60 GHz final hops, would require zero megahertz of 5 GHz band spectrum. The cable television companies could migrate their customers living in apartments to these configurations (see below regarding the importance of apartments). With these configurations, the cable television companies apparently could handle their projected spectrum shortage without asking for more spectrum. Instead, Wi-Fi providers seek large amounts of new spectrum, premised on the least spectrum-efficient configuration. Automobile companies, which innovate constantly to improve fuel efficiency and introduce automated vehicles and other technologies, are being asked to contribute their 5.9 GHz spectrum so that cable television companies can continue using the same Configuration A that has been used to deliver Wi-Fi service for the last nearly 20 years.
The focus should be on providing spectrum-efficient Wi-Fi service and not on allocating vast swathes of new spectrum for inefficient use. The Qualcomm Study makes good recommendations for implementation of spectrum-efficient Wi-Fi, including multiple AP installations per apartment and 60 GHz APs. The FCC should promote efficiency by premising any unlicensed spectrum allocation on greater Wi-Fi spectrum efficiency.
High band spectrum can satisfy much of the need currently projected for 5 GHz
The gating item in both studies, that drives up the amount of spectrum needed, is “frequency reuse which is needed to sustain the same target throughputs in neighboring apartments in the dense residential setting.”[40] To illustrate this point, the Qualcomm study calculates that a bungalow on the prairie (i.e., with no interference from overlapping networks) would need 80 megahertz of 5 GHz spectrum for four access points with Ethernet backhaul (Configuration B), while a similarly configured apartment in a multi-story apartment building would require 320 megahertz.[41] As noted above, the Wi-Fi Alliance Study similarly assumes a very dense residential setting (1076 square foot apartments, four residents per apartment, three devices per resident, multiple apartments per floor, multiple floors in the building). The 5 GHz radio signals from one apartment travel in all directions into neighboring apartments, interfering with the neighboring systems and requiring more spectrum so that everyone can use a channel that a neighbor does not already occupy. This geometrically drives up the amount of 5 GHz spectrum that is projected to be needed. The problem of overlapping Wi-Fi access points occurs most frequently in dense, high-rise apartment buildings and occurs less frequently, or not at all, in lower density settings like single family houses. Solving the frequency reuse problem in high-rise apartments would go a long way to resolving the projected shortage of unlicensed 5 GHz spectrum.
The two studies are also similar in that they assume that the projected shortfall in unlicensed spectrum must be made up by more mid-band spectrum, i.e., around 5 GHz. The Wi-Fi Alliance Study states that it chose “5 GHz rather than other bands because . . . many devices either cannot support 60 GHz or will not be within the range of a 60 GHz AP (access point).”[42] Similarly, for most scenarios, the Qualcomm Study assumes that signals on the last hop between the consumer’s device (smart phone, tablet, gaming console, etc.) and the access point(s) inside an apartment, travel over 5 GHz band spectrum. However, radio signals at 5 GHz have fairly good propagation characteristics, can travel a moderate distance, and can penetrate walls, floors and ceilings, to interfere with nearby systems. Thus, 5 GHz is not the best spectrum for a short hop in a single room in a high-rise apartment building.
By contrast, signals transmitted over high band (e.g., the 60 GHz band) spectrum travel shorter distances and do not propagate well through walls or windows.[43] High band spectrum is better suited to coverage in a single room where there are multiple access points in the residence. Indeed, the Qualcomm study states that “[l]arge scale use of 60 GHz technology for last hop links (between AP and STA) can significantly help mitigate some of the impacts of insufficient spectrum around the 5 GHz band.”[44] Qualcomm’s conclusion is reinforced by the study’s Configuration C, with the final hop traveling over the 60 GHz band rather than the 5 GHz band. As noted above, Configuration C would require just 480 megahertz of 5 GHz band spectrum – a significant reduction from the headline figure of 1280 megahertz needed, and less than the 580 megahertz that the FCC has already allocated.
Last hop links over high band spectrum are quite doable. Most high-rise apartment buildings (where the frequency reuse problem occurs) are wired with FIOS or Ethernet, which would facilitate more spectrum-efficient solutions. Further, there is an abundance of high band spectrum available. The FCC allocated, for unlicensed use, 14 gigahertz (14,000 megahertz) at 57-71 GHz (the “60 GHz band”). As the FCC noted, this 14 GHz is “15 times as much as all unlicensed spectrum for Wi-Fi in lower bands.”[45] Finally, the standard is in place. The IEEE 802.11ad (WiGig) standard makes feasible wireless local area networks (WLANs) and wireless personal area networks (WPANs) at 60 GHz. With a 2160-megahertz bandwidth at 60 GHz, WLANs and WPANs using WiGig can deliver gigabit Wi-Fi, and much higher speeds. What is needed is a regulatory and industry campaign to get devices to support 60 GHz, and to install multiple access points in high-rise apartments.
In conclusion, the 5.9 GHz band, that is allocated to DSRC, is simply not needed for a fulsome roll-out of high speed 5G and Wi-Fi. The FCC’s current allocation of unlicensed spectrum in the 5 GHz band, and pending allocation in the 6 GHz band, exceed reasonably projected demand for mid-band spectrum near 5 GHz, even where the spectrum-inefficient Configuration A is used. Deployment of devices using high band spectrum, as well as other spectrum-efficient configurations for apartments, should displace much projected 5 GHz usage. This should reduce the projected need for unlicensed mid band spectrum.
Second assumption: similar incumbents mean harmonized rules for unlicensed use
The 75 megahertz of spectrum that the Commission allocated to DSRC, at the 5.9 GHz band, sits immediately adjacent to the upper edge of the U-NII-3 band and also is immediately adjacent to the lower edge of the 6 MHz band that is under consideration in the FCC’s pending Mid BandProceeding discussed above.[46] If U-NII devices were to operate in the 5.9 GHz band under FCC rules identical to those in either the U-NII-3 band or the 6 MHz band, there presumably could be further opportunities to utilize the high-speed 80- and 160-megahertz bandwidths and to achieve economies of scale with uniform U-NII equipment.
Thus, in 2013, when it proposed that U-NII devices be allowed to share the 5.9 GHz band with incumbent radar and DSRC services, the FCC stated, “we believe that because the types of incumbent services across the 5 GHz spectrum share similar characteristics, the technical requirements for unlicensed devices also could share similar characteristics.”[47] In 2016, the Commission re-asserted its belief in similarity of incumbents and similarity of unlicensed devices.[48] However attractive this assumption may be on paper, it is not born out in fact.
Detect-and-avoid would impose stringent rules
The IEEE’s detect-and-avoid proposal would leave in place the DSRC band plan and require U-NII devices to vacate the 5.9 GHz band upon detection of a DSRC signal. Detect-and-avoid might possibly allow U-NII devices to co-exist without causing harmful interference to DSRC (as noted, the FCC’s report on Phase I testing has not been published as of this writing). However, detect-and-avoid would drive up the cost of U-NII devices and severely limit U-NII operation. U-NII devices could not apply their existing DFS protocols to the 5.9 GHz band in order to protect DSRC. DFS requires the U-NII device to vacate the channel within ten seconds of detecting radar. This ten second period is used mostly to identify an alternate channel and set up transmission on the new channel. By contrast, with the detect-and-avoid proposal, “there is no distinct channel move time; once a DSRC transmission is detected, unlicensed use of the band ceases immediately.”[49] Changing channels immediately requires different radio technology than identifying and setting up a new channel over a period of ten seconds. Further, detect-and-avoid would limit the time of transmission for any packet to three milliseconds,[50] while the FCC imposes no limit on packet length or time of transmission in the current U-NII bands,[51] and an 802.11ac packet is 5.46 milliseconds.[52] Finally, pervasive vehicle traffic generating DSRC signals in and around cities could effectively make the 5.9 GHz band not available to U-NII use in urban areas where Wi-Fi spectrum is most needed. Each DSRC-equipped vehicle will transmit a minimum of ten basic-safety-message radio signals per second, as well as additional vehicle to infrastructure radio signals. With just a few DSRC-equipped vehicles in an area, nearby U-NII devices could be effectively foreclosed from using the 5.9 GHz band. In urban areas, vehicular traffic is pervasive. By contrast, the government radar installations, that generate signals protected by DFS, are not as prevalent as vehicular traffic. In many areas, the DFS-limited channels are clear of radar. These and other differences make detect-and-avoid more stringent than DFS.
U-NII operators have found the DFS-limited spectrum bands to be less useful than other bands. Consequently, default settings on many U-NII devices automatically exclude the DFS-limited channels, or the operator will manually program the device to avoid the DFS-limited channels. In comments to the FCC, the Wi-Fi Alliance summarized its concerns with DFS:
The following are among the issues contributing to lighter use of the bands in which DFS requirements are imposed: (1) as noted below, DFS limits the use of some applications; (2) additional time is needed to obtain equipment approval for products using DFS frequencies; and (3) there is added complexity in designing and producing equipment with DFS capabilities. In addition, once successfully engineered into devices, DFS creates an ongoing operational cost for the device or system due to regulatory obligations for channel availability check time, in-service monitoring, and non-occupancy period, among other requirements. These requirements make a system operating on DFS-limited channels less available than one that uses only unencumbered channels, particularly if the system detects false radar patterns.[53]
If the DFS-limited spectrum bands are unattractive, the 5.9 GHz band would be especially unattractive to U-NII operators. All of the things the Wi-Fi Alliance said about DFS would apply with greater force to the 5.9 GHz band if subjected to detect-and-avoid rules. Detect-and-avoid would limit the use of U-NII applications, would require additional time for U-NII equipment approval, would add complexity to the design and production of U-NII equipment and would impose ongoing operational costs. As the IEEE Tiger Team noted with respect to detect-and-avoid, “[c]hanges will be required to modify the behavior of existing 802.11ac systems . . . [t]his would require changes in the base 802.11 specification and would add complexity to existing 802.11 chipsets.”[54] Re-designed and more complex U-NII equipment undoubtedly would be more expensive than current equipment. Finally, U-NII systems operating on the 5.9 GHz band would be less available than U-NII systems that use only unencumbered channels.
As noted, the 5.9 GHz band is immediately adjacent to the U-NII-3 band on one side and the 6 GHz band tentatively slated for unlicensed use, on the other side. Ordinarily, with uniform rules for unlicensed use, this spectrum could be aggregated to form valuable 80- and 160-megahertz channels. However, with the 5.9 GHz band subject to detect-and-avoid rules, these wider channels could not be readily formed. The only way to harmonize rules would be to subject the surrounding spectrum to the much more restrictive detect-and-avoid rules of the 5.9 GHz band when setting up 80- or 160-megahertz channels. That is not an attractive option for U-NII operators.
Third Assumption: President Obama’s Memorandum and the Spectrum Act impose policy and legal requirements
In 2013, when it proposed 5.9 GHz spectrum-sharing, the FCC’s NPRM presented little or no evidence of need for the 5.9 GHz band to be allocated to unlicensed use.[55] However, the NPRM cited to President Obama’s Presidential Memorandum of June 2010, which called for the National Telecommunications and Information Administration (NTIA) and the FCC to make available a total of 500 megahertz of licensed and unlicensed spectrum for commercial mobile and fixed wireless broadband use by the year 2020.[56] In the intervening years since issuance of this Presidential Memorandum, the FCC has done an admirable job of allocating spectrum for broadband and has more than satisfied President Obama’s goal. These allocations include the AWS-3 Auction (making available 65 megahertz in the 1695-1710 MHz, 1755-1780 MHz and 2155-2180 MHz bands),[57] the 3.5 GHz R&O (making available 150 megahertz in the 3550-3700 MHz band),[58] the Incentive Auction (making available 84 megahertz of spectrum at 600 MHz),[59] the Mid-Band Spectrum Proceeding (inquiring about making available 1700 megahertz of spectrum in the 3.7-4.2 GHz and 5.925 -7.125 GHz bands),[60] and the Spectrum Frontiers Proceeding.[61] In the Spectrum Frontiers Proceeding, the Commission made available 10.85 gigahertz (10,850 megahertz) of spectrum at various bands above 28 GHz, including seven gigahertz (7000 megahertz) of unlicensed spectrum at 64-71 GHz, and proposed to make available an additional 18 gigahertz of high-band spectrum.[62] In a follow-on order in the Spectrum Frontiers Proceeding, the Commission made available an additional 1700 megahertz of high-band spectrum.[63] In short, the FCC has exceeded President Obama’s spectrum allocation goal.
The NPRM also cites to the Spectrum Act,[64] which required the Commission to allow unlicensed U-NII devices to operate in the U-NII-2B band if it determines, in consultation with NTIA, that Federal users would be protected.[65] The Commission has discharged its duties under the Spectrum Act by considering whether to allow unlicensed devices in the U-NII-2B band. As then-Commissioner Pai noted, the NPRM’s proposal to also allocate the 5.9 GHz band to U-NII devices went beyond what was required by the Spectrum Act.[66] In conclusion, neither President Obama’s Memorandum nor the Spectrum Act requires the FCC to share the 5.9 GHz band with U-NII devices.
References
[1] U.S. Department of Transportation. (July 2015). Status of the Dedicated Short-Range Communications Technology and Applications, Report to Congress, Appendix J, Executive Summary, FHWA-JPO-15-218.
[2] Barker, J. (2014). Letter to Marlene Dortch, Federal Communications Commission, ET Docket No. 13-49.
[3] Federal Communications Commission. (1999). In the Matter of Amendment of Parts 2 and 90 of the Commission’s Rules to Allocate the 5.850-5.935 GHz Band to the Mobile Service for Dedicated Short Range Communications of Intelligent Transportation Services, Report and Order, ET Docket No. 98-9514.
[4] Federal Communications Commission. (2013). In the Matter of Revision of Part 15 of the Commission’s Rules to Permit Unlicensed National Information Infrastructure Devices in the 5 GHz Band, Notice of Proposed Rulemaking, ET Docket No. 13-49 (“NPRM”).
[5] The Institute of Electrical & Electronic Engineers. (2015). Final Report of the DSRC Coexistence Tiger Team, doc.: IEEE 802.11-15/0347r0 (“Tiger Team Final Report”).
[6] Federal Communications Commission. (2016). The Commission Seeks to Update and Refresh the Record in the “National Information Infrastructure (U-NII) Devices in the 5 GHz Band” Proceeding, Public Notice, ET Docket No. 13-49.
[7] Id page 11.
[8] Motorola, Inc. (2002). Canopy Wireless Broadband Internet Platform, Configuration Guide, page 5, http://winncom.com/pdf/MotorolaConfigurationGuide.pdf (accessed Feb 26, 2018).
[9] Federal Communications Commission. (2014). In the Matter of Revision of Part 15 of the Commission’s Rules to Permit Unlicensed National Information Infrastructure (U-NII) Devices in the 5 GHz Band, First Report and Order, ET Docket. No. 13-49 (“First R&O”).
[10] First R&O, para. 2.
[11] NPRM, para. 4.
[12] 47 C.F.R. § 15.407(h).
[13] Id.
[14] 47 C.F.R. § 15.407.
[15] First R&O, para. 24.
[16] Consolidated Appropriations Act, 2018, P.L. 115-142, 115th Congress, Second Session, March 23, 2018.
[17] See e.g., Federal Communications Commission. (2014). Expanding the Economic and Innovation Opportunities of Spectrum Through Incentive Auctions, Report and Order, GN Docket No. 12-268, pages 1-3.
[18] Federal Communications Commission. (1997). FCC Report to Congress On Spectrum Auctions, pages 6-13, http://wireless.fcc.gov/auctions/data/papersAndStudies/fc970353.pdf (accessed Feb 26, 2018).
[19] Federal Communications Commission. (2017). Wireless Telecommunications Bureau Reminds Wireless Licensees of Construction Obligations, Public Notice.
[20] Quotient Associates. (2017). “Wi-Fi Spectrum Needs Study, Final Report to Wi-Fi Alliance” http://www.wi-fi.org/file/wi-fi-spectrum-needs-study (accessed Feb. 26, 2018) (“Wi-Fi Alliance Study”).
[21] Rolf De Vegt, et al., Qualcomm Incorporated. (2016). “A Quantification of 5 GHz Unlicensed BandSpectrum Needs” https://www.qualcomm.com/documents/quantification-5-ghz-unlicensed-band-spectrum-needs, (accessed Feb 26, 2018) (“Qualcomm Study”).
[22] Wi-Fi Alliance Study, page 5.
[23] Wi-Fi Alliance Study, page 1.
[24] Wi-Fi Alliance Study, pages 6-7.
[25] Wi-Fi Alliance Study, pages 6-7.
[26] Wi-Fi Alliance Study, page 16.
[27] Wi-Fi Alliance Study, page 6.
[28] Wi-Fi Alliance Study, page 31.
[29] Wi-Fi Alliance Study, pages 23-24.
[30] Wi-Fi Alliance Study, page 24.
[31] Wi-Fi Alliance Study, page 21.
[32] Wi-Fi Alliance Study, page 26.
[33] Id.
[34] Qualcomm Study, page 5.
[35] Qualcomm Study, page 6.
[36] Qualcomm Study, page 4.
[37] Federal Communications Commission. (2017). In the Matter of Expanding Flexible Use in Mid-Band Spectrum Between 3.7 and 24 GHz, Notice of Inquiry, GN Docket No. 17-183 (the “Mid-Band Spectrum Proceeding”).
[38] O’Rielly, M. (2017). A Mid-Band Spectrum Win in the Making, FCC Blog of Commissioner Michael O’Rielly, https://www.fcc.gov/news-events/blog/2017/07/10/mid-band-spectrum-win-making (accessed Sep 21, 2017).
[39] IEEE802 LAN/MAN Standards Committee. (2017). “Comments of IEEE802” In the Matter of Expanding Flexible Use in Mid-Band Spectrum Between 3.7 and 24 GHz, Federal Communications Commission GN Docket No. 17-183.
[40] Qualcomm Study, page 9.
[41] Qualcomm Study, page 7.
[42] Wi-Fi Alliance Study, page 17.
[43] Smulders, P.F.M., L.M. Corriea (1997). Characterisation of propagation in 60 GHz radio channels, Electronics & Communication Engineering Journal, vol. 9, issue 2.
[44] Qualcomm Study, page 19.
[45] Federal Communications Commission. (2016). Fact Sheet: Spectrum Frontiers Proposal to Identify, Open Up Vase Amounts of New High-Band Spectrum For Next Generation (5G) Wireless Broadband. https://apps.fcc.gov/edocs_public/attachmatch/DOC-339990A1.pdf
[46] Federal Communications Commission. (2017). In the Matter of Expanding Flexible Use in Mid-Band Spectrum Between 3.7 and 24 GHz, Notice of Inquiry, GN Docket No. 17-183 (the “Mid-Band Spectrum Proceeding”).
[47] NPRM, para. 95.
[48] Federal Communications Commission. (2016). The Commission Seeks to Update and Refresh the Record in the “National Information Infrastructure (U-NII) Devices in the 5 GHz Band” Proceeding, Public Notice, ET Docket No. 13-49.
[49] Tiger Team Final Report, page 6.
[50] Id.
[51] 47 C.F.R. § 15.407.
[52] Rohde & Schwarz. (2012). “802.11ac Technology Introduction White Paper” page 14 https://cdn.rohde-schwarz.com/pws/dl_downloads/dl_application/application_notes/1ma192/1MA192_7e_80211ac_technology.pdf (accessed March 7, 2018).
[53] Wi-Fi Alliance. (2017). Comments of Wi-Fi Alliance, Federal Communications Commission GN Docket No. 17-183.
[54] Tiger Team Final Report, page 7.
[55] NPRM, paras. 11-13.
[56] NPRM at para 11, citing President Barak Obama. (2010). Presidential Memorandum: Unleashing the Wireless Broadband Revolution (Executive Memo)(June 28, 2010), https://obamawhitehouse.archives.gov/the-press-office/presidential-memorandum-unleashing-wireless-broadband-revolution(accessed Feb. 26, 2018).
[57] Federal Communications Commission. (2014). In the Matter of Amendment of the Commission’s Rules With Regard to Commercial Operations in the 1695-1710 MHz, 1755-1780 MHz and 2155-2180 MHz Bands, Report and Order, GN Docket No. 13-185 (the “AWS-3 Auction”).
[58] Federal Communications Commission. (2015). In the Matter of Amendment of the Commission’s Rules With Regard to Commercial Operations in the 3550-3650 MHz Band, Report and Order and Further Notice of Proposed Rulemaking, GN Docket No. 12-354 (“3.5 GHz R&O”).
[59] Federal Communications Commission. (2014). In the Matter of Expanding the Economic and Innovation Opportunities of Spectrum Through Incentive Auctions, Report and Order, GN Docket No. 12-268 (the “Incentive Auction”).
[60] Federal Communications Commission. (2017). In the Matter of Expanding Flexible Use in Mid-Band Spectrum Between 3.7 and 24 GHz, Notice of Inquiry, GN Docket No. 17-183 (the “Mid-Band Spectrum Proceeding”).
[61] Federal Communications Commission. (2016). In the Matter of Use of Spectrum Bands Above 24 GHz for Mobile Radio Services, Report and Order and Further Notice of Proposed Rulemaking, GN Docket No. 14-177 (the “Spectrum Frontiers Proceeding”).
[62] Id.
[63] Federal Communications Commission. (2017). In the Matter of Use of Spectrum Bands Above 24 GHz for Mobile Radio Services, Second Report and Order, Second Further Notice of Proposed Rulemaking, Order on Reconsideration, and Memorandum Opinion and Order, FCC17-152, GN Docket No 14-177.
[64] NPRM para 12.
[65] Middle Class Tax Relief and Job Creation Act of 2012, Pb. L. No. 112-96, § 6406, 126 Stat. 156, 231 (2012), 47 U.S.C. § 1453 (Spectrum Act).
[66] Pai, A. (2013). Statement of Commissioner Ajit Pai, In the Matter of Revision of Part 15 of the Commission’s Rules to Permit Unlicensed National Information Infrastructure Devices in the 5 GHz Band, Notice of Proposed Rulemaking, ET Docket No. 13-49